Summary Significant progress has been made in using neuroimaging and cerebrospinal fluid (CSF) protein measurements as Alzheimer disease (AD) biomarkers. However, there is still a critical need to identify more reliable and reproducible biomarkers with further improved diagnostic accuracy, especially to differentiate AD from other dementias, to track or monitor the disease progression and to objectively evaluate drug effects. There is also a growing interest in developing novel biomarkers that could reflect different aspects of AD pathology and accurately detect pathogenic components of AD before appearance of significant cognitive decline, thereby assisting with AD diagnosis at early symptomatic and even preclinical stages. This proposal is designed to meet several major challenges of current biomarker research, specifically: 1) difficulties in development of antibody-based, quantitative protein assays for most novel candidates identified by proteomic profiling and significant variations associated with most existing immunoassays, 2) low sensitivity and specificity of blood-based markers, and 3) detection of AD at early or even preclinical stages. To address the problems of antibody-based assays, our strategy is development of targeted mass spectrometry-based techniques, such as selected reaction monitoring (SRM), to identify unique peptide markers derived from proteins either showing promise in previous proteomics profiling, or known to be critical to AD pathogenesis in human cerebrospinal fluid (CSF). To facilitate discovery and validation of blood based biomarkers, a specific population of central nervous system derived plasma exosomes, the cargo-carrying microvesicles recognized recently to transport biomolecules among different cells or organ systems, will be isolated before SRM analysis. The unique peptide markers will be tested in several large, well-established cohorts, e.g., Alzheimer's Disease Research Centers (ADRCs) affiliated with the University of Washington, Oregon Health and Science University, University of California at San Diego, and University of Pennsylvania, and ADNI (Alzheimer's Disease Neuroimaging Initiative), with cross-sectional and longitudinal samples collected, along with extensive clinical characterization. Finally, to improve early diagnosis, we will make use of a very early MCI cohort consisting of subjects at elevated risk for AD, with the goal of discovering biomarkers capable of identifying subjects with early or preclinical AD. The studies designed for this project, if successful, have the potential to result in a panel(s) of biomarkers that are robust, with less variation than can currently be achieved, and in a body fluid that is readily accessible in a regular clinical setting. Markers for early diagnosis and progression of AD are critical in understanding how to arrest or slow AD progression.